MatchLab AI: Ensuring High-Quality Clinical Imaging for Better Treatment
healthcaretechoutlook

MatchLab AI: Ensuring High-Quality Clinical Imaging for Better Treatment

Follow MatchLab AI on :

Nick Rance, CEO, MatchLab AINick Rance, CEO, MatchLab AI
The successful journey of MatchLab AI dates back to the team initially meeting at MIT’s Hacking Dermatology event. The team was looking for ways to build an AI that could help dermatologists monitor a patient’s disease over time without the need to consult a doctor. While working on this AI they realized that there was a problem further upstream in the data pipeline, they had observed the struggle that dermatologists went through to collect high-quality standardized pictures of their patient’s skin for further reference. The images that dermatologists often receive are out of focus, over or underexposed, taken at awkward angles and, most importantly, difficult to compare - resulting in ambiguity when tracking treatment progress. Recognizing this foundational problem, the team took matters into their own hands and built an application that enabled everyone, even a layman, to capture clinical-grade photos of any area of their body using their smartphone. This was the genesis of MatchLab AI.

Today, MatchLab AI uses their breakthroughs in computer vision to power the simplest, smartest app in dermatology. “We have developed an algorithm that automatically enables a camera to lock its focus on a relevant rash or lesion and ensure that the picture will be clinically helpful and longitudinally comparable,” states Nick Rance, the software engineer turned CEO of MatchLab AI. But that was just one part of the roadblock that MatchLab AI had managed to overcome. They still needed to surmount other challenges such as setting up an appropriate lighting feature for the photos that were taken.


We want our application to be as seamlessly integrated and intuitive as possible - the first piece of medical software that doctors enjoy using


The team was examining data from various patient support groups where they observed that most of the pictures of the lesions were taken in dimly lit areas such as bathrooms lowering the visibility of the photos. This prompted them to implement an image standardization technique in the application so that the patient could only capture an image after meeting a certain threshold point in lighting.

The company’s distinctive AI-based image capturing application guides patients to recreate the same pose that they did previously through an overlay network. The solution identifies the user’s position and makes an outline based on the data obtained. The user matches their body pose against the outline and the camera automatically clicks the photo after making sure that the portion corresponds to the same position. The pictures can be transformed into the form of a gif for the doctors to quickly evaluate a condition over a period of time and decide the disease’s progress or improvement. These gifs, photos and other useful documentation will be available to dermatologists through an easy to use portal which analyzes patient data and assists them in treating their patients.

In a nutshell, the team has developed a groundbreaking application that allows anyone to capture clinical-grade photos of any area of the body with the touch of a button.

For the road ahead, MatchLab AI is all set to hit the medical market over the course of 2019 through partnerships with major research institutions and private practices. Immediately, MatchLab AI intends to help dermatologists and HCPs at hospitals and in private practice and, over the next 4 years, empower all patients to take control of their own skin health. “We want our application to be as seamlessly integrated and intuitive as possible - the first piece of medical software that doctors enjoy using!” remarks Rance.

Share this Article:
MatchLab AI

Company
MatchLab AI

Headquarters
Cambridge, MA

Management
Nick Rance, CEO

Description
MatchLab AI uses the latest breakthroughs in computer vision to power the simplest app in...